Sorry there have been a number of interruptions to the SOTA Cluster. This has been due to me having to try to program around some new users, who through ignorance, are abusing the system. I fear the work of AI assistance here.
I’ve mailed the US station who managed to get over 30 simultaneous connections but wasn’t actually looking at the results obtained. He’s been emailed but is close to being firewalled. There’s another Swiss station running bots about to be mailed about poor connection practice and likewise a French station I’m still looking for contact details.
It’s not malicious, just poor understanding and implementation. And, TBH, a bit of stupidity in people setting up automatic systems and never looking at the results and adjusting what they are doing.
And we have the German Telekom customer who has been trying to connect for 5 months now. Obviously he (it will not be a she, ladies are too smart) has forgotten he is trying to connect. I contacted abuse@telekom.de who were as much use as a chocolate fireguard wanting me to install their app, they can’t understand I’m not a customer. Sadly it’s a dynamic IP and I’m reaching the point of firewalling 1.7million telekom.de users. A bit extreme but I’m getting annoyed now.
The US station has been in touch and apologised for his clumsiness, no issues with him now. It’s nice for someone to hold up their hand and say sorry, appreciated.
No reply from HB9. Into the firewall he went along with some others doing the same strange connect/disconnect/reconnect pattern. I fear this is some new app that’s out because 1 person does this then this is odd, 2 doing it is strange and 3 or more is a movement (to paraphrase Alice’s Restaurant). What makes me sad is that 3 people went into the firewall and they’re still banging away trying to do something and nobody is looking at why it’s not working, stopping whatever is running and maybe asking me a question or two. My curmudgeonly like personality just knows this is some blige AI app at the core.
The cluster is a little bit fragile at present. There has been a large uptake in use and this has pushed me into fixing a few minor issues. One was that no matter what you asked for you got debug logging and with more users that was writing lots to the disk. The other was a memory leak. Both of those are fixed but fixing the leak has introduce a thread safe access issue. I’ve tried locating this in a test version but there’s not enough traffic. So I’m doing a Microsoft and testing in on my users. Sorry! This means the cluster may restart for no good reason or disappear offline for some minutes. I’ll post here when I find the cause.
Well the latest version of software has been running for 985mins so far with no issues. I don’t like my fix but at least we have a cluster for people to use. I’ll continue breaking er no, enhancing it
